Abstract
We describe an algorithm for the detection of ellipse shapes in images, using the Randomized Hough Transform. The method is compared with three other Hough-based algorithms. Tests are performed using both noise-free and noisy images, and several real-world images. The algorithm was found to give improvements in accuracy, and a reduction in computation time and the number of false alarms detected. A program allowing the reader to experiment with these algorithms can be found at the WWW address http://ciips.ee.uwa.edu.au/Papers/Journal_Papers/1998/01/Index.html. (C) 1998 Elsevier Science B.V. All rights reserved.
Original language | English |
---|---|
Pages (from-to) | 299-305 |
Journal | Pattern Recognition Letters |
Volume | 19 |
Issue number | 3-4 |
DOIs | |
Publication status | Published - 1998 |